Principles of a good data mining report
I have written quite a lot of reports, highlighting results from a wide range of activities. Based on the feedback received and from the requests of changes intervened from the first version to the definitive one, I can tell you the following basic principles for producing a good data mining report:
- Clearly show the objectives and questions that initiated the analyses performed
- Explicitly highlight the assumptions made when performing the analyses
- Enumerate and get into the details with data treatment applied to analyzed data
- Always verify that the reproduced data is consistent
- Provide data lineage to the maximum possible extent
Let's get a bit deeper into these principles before we actually develop our report.
Clearly state the objectives
When I was a newbie in the field, I got really excited at the idea of applying all the data mining techniques I knew by then to real data. You can imagine my happiness when I was first provided with a bunch of data. I was...